Keras: Input 0 of layer "lstm" is incompatible with the layer: expected ndim=3, found ndim=2

Question

I'm making a paraphrase detection model. My current model is as follows:

left_input = Input(shape=(120, ))
right_input = Input(shape=(120, ))
left_embedding = Embedding(vocab_size, 120, input_length=max_length)(left_input)
right_embedding = Embedding(vocab_size, 120, input_length=max_length)(right_input)
left_lstm = LSTM(120, )(left_embedding)
right_lstm = LSTM(120, )(right_embedding)
concat = concatenate([left_lstm, right_lstm], name='Concatenate')
model_output = Dense(2, activation='softmax')(concat)

model = Model(inputs=[left_embedding, right_embedding], outputs=model_output, name='Final_output')
model.compile(optimizer='adam', loss='binary_crossentropy')

But when I try to fit, it throws an error:

model.fit(x=[train_sentences_1, train_sentences_2], y=train_labels,
            validation_data=([val_sentences_1, val_sentences_2], val_labels), 
            epochs = 10, batch_size = 32)

Input 0 of layer "lstm" is incompatible with the layer: expected ndim=3, found ndim=2. Full shape received: (None, 120).

train_sentences_1, train_sentences_2 are of shape (10000, 120) while train_labels is of shape (10000, 1)

Can someone help me in fixing the problem? I understand its complaining about dimensionality error in LSTM error but I'm not sure how it can be fixed.

Thanks!